The chart above plots trends over time in successful coups worldwide according to two prominent data sets: Center for Systemic Peace and Powell & Thyne.
- The dots mark the annual counts observed by the two sources. If you only see one dot for a particular year, it means the two sources saw the same number of events, so the green dot is covering the brown one.
- The lines are locally weighted regression estimates (loess) of the time trends those dots describe. Think of the annual counts as noisy indicators of an underlying process—-coup risk—-that we’re really interested in but can’t directly observe. The lines represent statistical estimates of that underlying process.
Moral of the story: It might feel like coups are happening more often of late, but that’s not what the data show so far.
PS These are raw counts, and the number of countries worldwide has increased significantly since the 1980s. If we used the rate instead of the count, the decline over the past 30 years would look even more dramatic.
PPS The same basic time trend holds for failed coups, too.